aboutsummaryrefslogtreecommitdiff
path: root/lib/Transforms
AgeCommit message (Collapse)Author
2009-02-05Ignore dbg intrinsics while folding switch instruction.Devang Patel
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63802 91177308-0d34-0410-b5e6-96231b3b80d8
2009-02-04Ignore dbg intrinsics.Devang Patel
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63781 91177308-0d34-0410-b5e6-96231b3b80d8
2009-02-04While folding vallue comparison terminators ignore dbg intrinsics.Devang Patel
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63700 91177308-0d34-0410-b5e6-96231b3b80d8
2009-02-04Ignore dbg intrinsics while hoisting common code in the two blocks up into ↵Devang Patel
the branch block. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63687 91177308-0d34-0410-b5e6-96231b3b80d8
2009-02-03Do not let dbg intrinsic block folding of two entry phi node.Devang Patel
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63671 91177308-0d34-0410-b5e6-96231b3b80d8
2009-02-03If "optimize for size" attribute is set then block non-trivial loop ↵Devang Patel
unswitches but allow trivial loop unswitches. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63670 91177308-0d34-0410-b5e6-96231b3b80d8
2009-02-03teach "convert from scalar" to handle loads of fca's.Chris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63659 91177308-0d34-0410-b5e6-96231b3b80d8
2009-02-03refactor the interface to ConvertUsesOfLoadToScalar, Chris Lattner
renaming it to ConvertScalar_ExtractValue git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63658 91177308-0d34-0410-b5e6-96231b3b80d8
2009-02-03convert ConvertUsesOfLoadToScalar to use IRBuilder,Chris Lattner
no functionality change. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63652 91177308-0d34-0410-b5e6-96231b3b80d8
2009-02-03switch ConvertScalar_InsertValue to use an IRBuilder, no Chris Lattner
functionality change. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63651 91177308-0d34-0410-b5e6-96231b3b80d8
2009-02-03make scalar conversion handle stores of first classChris Lattner
aggregate values. loads are not yet handled (coming soon to an sroa near you). git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63649 91177308-0d34-0410-b5e6-96231b3b80d8
2009-02-03Make SROA produce a vector only when the alloca is actually Chris Lattner
accessed at least once as a vector. This prevents it from compiling the example in not-a-vector into: define double @test(double %A, double %B) { %tmp4 = insertelement <7 x double> undef, double %A, i32 0 %tmp = insertelement <7 x double> %tmp4, double %B, i32 4 %tmp2 = extractelement <7 x double> %tmp, i32 4 ret double %tmp2 } instead, producing the integer code. Producing vectors when they aren't otherwise in the program is dangerous because a lot of other code treats them carefully and doesn't want to break them down. OTOH, many things want to break down tasty i448's. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63638 91177308-0d34-0410-b5e6-96231b3b80d8
2009-02-03APInt'fy SimplifyDemandedVectorElts so it can analyze vectors with more than ↵Evan Cheng
64 elements. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63631 91177308-0d34-0410-b5e6-96231b3b80d8
2009-02-03add another case of undefined behavior without crashing, PR3466.Chris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63620 91177308-0d34-0410-b5e6-96231b3b80d8
2009-02-03Revert r63600. It didn't fix the bug, it just moved it a bit.Nick Lewycky
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63618 91177308-0d34-0410-b5e6-96231b3b80d8
2009-02-03Update the callgraph when replacing InvokeInst with CallInst when inlining.Nick Lewycky
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63600 91177308-0d34-0410-b5e6-96231b3b80d8
2009-02-03Teach ConvertUsesToScalar to handle memset, allowing it to handle Chris Lattner
crazy cases like: struct f { int A, B, C, D, E, F; }; short test4() { struct f A; A.A = 1; memset(&A.B, 2, 12); return A.C; } git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63596 91177308-0d34-0410-b5e6-96231b3b80d8
2009-02-03rearrange how SRoA handles promotion of allocas to vectors.Chris Lattner
With the new world order, it can handle cases where the first store into the alloca is an element of the vector, instead of requiring the first analyzed store to have the vector type itself. This allows us to un-xfail test/CodeGen/X86/vec_ins_extract.ll. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63590 91177308-0d34-0410-b5e6-96231b3b80d8
2009-02-02inline SROA::ConvertToScalar, no functionality change.Chris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63544 91177308-0d34-0410-b5e6-96231b3b80d8
2009-02-02Fix a bug which caused us to miscompile a couple of AdaChris Lattner
tests. Thanks for the beautiful reduced testcase Duncan! git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63529 91177308-0d34-0410-b5e6-96231b3b80d8
2009-02-02Fix a comment (bytes -> bits), reformat a commentDuncan Sands
and remove trailing whitespace. No functionality change. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63511 91177308-0d34-0410-b5e6-96231b3b80d8
2009-02-02Fix an obvious thinko.Duncan Sands
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63510 91177308-0d34-0410-b5e6-96231b3b80d8
2009-02-02reduce indentation, (~XorCST->getValue()).isSignBit() -> isMaxSignedValue()Chris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63500 91177308-0d34-0410-b5e6-96231b3b80d8
2009-01-31Reinstate this optimization to fold icmp of xor when possible. Don't try toNick Lewycky
turn icmp eq a+x, b+x into icmp eq a, b if a+x or b+x has other uses. This may have been increasing register pressure leading to the bzip2 slowdown. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63487 91177308-0d34-0410-b5e6-96231b3b80d8
2009-01-31Fix PR3452 (an infinite loop bootstrapping) by disabling the recentChris Lattner
improvements to the EvaluateInDifferentType code. This code works by just inserted a bunch of new code and then seeing if it is useful. Instcombine is not allowed to do this: it can only insert new code if it is useful, and only when it is converging to a more canonical fixed point. Now that we iterate when DCE makes progress, this causes an infinite loop when the code ends up not being used. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63483 91177308-0d34-0410-b5e6-96231b3b80d8
2009-01-31now that all the pieces are in place, teach instcombine'sChris Lattner
simplifydemandedbits to simplify instructions with *multiple uses* in contexts where it can get away with it. This allows it to simplify the code in multi-use-or.ll into a single 'add double'. This change is particularly interesting because it will cover up for some common codegen bugs with large integers created due to the recent SROA patch. When working on fixing those bugs, this should be disabled. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63481 91177308-0d34-0410-b5e6-96231b3b80d8
2009-01-31simplify/clarify control flow and improve comments, no functionality change.Chris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63480 91177308-0d34-0410-b5e6-96231b3b80d8
2009-01-31make some fairly meaty internal changes to how SimplifyDemandedBits works.Chris Lattner
Now, if it detects that "V" is the same as some other value, SimplifyDemandedBits returns the new value instead of RAUW'ing it immediately. This has two benefits: 1) simpler code in the recursive SimplifyDemandedBits routine. 2) it allows future fun stuff in instcombine where an operation has multiple uses and can be simplified in one context, but not all. #2 isn't implemented yet, this patch should have no functionality change. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63479 91177308-0d34-0410-b5e6-96231b3b80d8
2009-01-31minor cleanupsChris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63477 91177308-0d34-0410-b5e6-96231b3b80d8
2009-01-31make sure to set Changed=true when instcombine hacks on the code,Chris Lattner
not doing so prevents it from properly iterating and prevents it from deleting the entire body of dce-iterate.ll git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63476 91177308-0d34-0410-b5e6-96231b3b80d8
2009-01-31Simplify and generalize the SROA "convert to scalar" transformation toChris Lattner
be able to handle *ANY* alloca that is poked by loads and stores of bitcasts and GEPs with constant offsets. Before the code had a number of annoying limitations and caused it to miss cases such as storing into holes in structs and complex casts (as in bitfield-sroa) where we had unions of bitfields etc. This also handles a number of important cases that are exposed due to the ABI lowering stuff we do to pass stuff by value. One case that is pretty great is that we compile 2006-11-07-InvalidArrayPromote.ll into: define i32 @func(<4 x float> %v0, <4 x float> %v1) nounwind { %tmp10 = call <4 x i32> @llvm.x86.sse2.cvttps2dq(<4 x float> %v1) %tmp105 = bitcast <4 x i32> %tmp10 to i128 %tmp1056 = zext i128 %tmp105 to i256 %tmp.upgrd.43 = lshr i256 %tmp1056, 96 %tmp.upgrd.44 = trunc i256 %tmp.upgrd.43 to i32 ret i32 %tmp.upgrd.44 } which turns into: _func: subl $28, %esp cvttps2dq %xmm1, %xmm0 movaps %xmm0, (%esp) movl 12(%esp), %eax addl $28, %esp ret Which is pretty good code all things considering :). One effect of this is that SROA will start generating arbitrary bitwidth integers that are a multiple of 8 bits. In the case above, we got a 256 bit integer, but the codegen guys assure me that it can handle the simple and/or/shift/zext stuff that we're doing on these operations. This addresses rdar://6532315 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63469 91177308-0d34-0410-b5e6-96231b3b80d8
2009-01-30use precise gettersGabor Greif
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63402 91177308-0d34-0410-b5e6-96231b3b80d8
2009-01-28Fix some issues with volatility, move "CanConvertToScalar" check Chris Lattner
after the others. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63227 91177308-0d34-0410-b5e6-96231b3b80d8
2009-01-28Rename getAnalysisToUpdate to getAnalysisIfAvailable.Duncan Sands
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@63198 91177308-0d34-0410-b5e6-96231b3b80d8
2009-01-26Fixed optimization of combining two shuffles where the first shuffle inputsMon P Wang
has a different number of elements than the output. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62998 91177308-0d34-0410-b5e6-96231b3b80d8
2009-01-26Handle single-entry phi nodes gracefully in condprop.Chris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62985 91177308-0d34-0410-b5e6-96231b3b80d8
2009-01-26Fix PR3408 by making a non-obvious assumption very obvious, and Chris Lattner
handling the flaw inherent in that assumption. :) git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62984 91177308-0d34-0410-b5e6-96231b3b80d8
2009-01-26More cleanups and simplifications, no functionality change.Chris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62983 91177308-0d34-0410-b5e6-96231b3b80d8
2009-01-26tidy assertsChris Lattner
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62982 91177308-0d34-0410-b5e6-96231b3b80d8
2009-01-25The function that does nothing but call malloc is noalias return.Nick Lewycky
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62956 91177308-0d34-0410-b5e6-96231b3b80d8
2009-01-24Revert previous change; even this mild and clearlyDale Johannesen
more accurate change loses more than it gains on benchmarks. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62938 91177308-0d34-0410-b5e6-96231b3b80d8
2009-01-24testcase for PR3381.Torok Edwin
Also it was an empty struct, not a void after all. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62920 91177308-0d34-0410-b5e6-96231b3b80d8
2009-01-24void* is represented as pointer to empty struct {}.Torok Edwin
Thus we need to check whether the struct is empty before trying to index into it. This fixes PR3381. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62918 91177308-0d34-0410-b5e6-96231b3b80d8
2009-01-24Improve the inlining cost function a bit.Dale Johannesen
Little practical effect. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62908 91177308-0d34-0410-b5e6-96231b3b80d8
2009-01-24Make InstCombineStoreToCast handle aggregates more aggressively,Chris Lattner
handling the case in Transforms/InstCombine/cast-store-gep.ll, which is a heavily reduced testcase from Clang on x86-64. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62904 91177308-0d34-0410-b5e6-96231b3b80d8
2009-01-23use CallSite::isCalle instead of slow getOperandNoGabor Greif
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62877 91177308-0d34-0410-b5e6-96231b3b80d8
2009-01-23Simplify the logic of getting hold of a PHI predecessor block.Gabor Greif
There is now a direct way from value-use-iterator to incoming block in PHINode's API. This way we avoid the iterator->index->iterator trip, and especially the costly getOperandNo() invocation. Additionally there is now an assertion that the iterator really refers to one of the PHI's Uses. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62869 91177308-0d34-0410-b5e6-96231b3b80d8
2009-01-22introduce a useful abstraction to find out if a Use is in the call position ↵Gabor Greif
of an instruction git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62788 91177308-0d34-0410-b5e6-96231b3b80d8
2009-01-21Remove uses of uint32_t in favor of 'unsigned' for better Chris Lattner
compatibility with cygwin. Patch by Jay Foad! git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62695 91177308-0d34-0410-b5e6-96231b3b80d8
2009-01-21Make special cases (0 inf nan) work for frem.Dale Johannesen
Besides APFloat, this involved removing code from two places that thought they knew the result of frem(0., x) but were wrong. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@62645 91177308-0d34-0410-b5e6-96231b3b80d8
generated by cgit v1.2.3-18-g5258 (git 2.32.0) at 2025-07-19 22:17:25 +0000